Anthropic launched Claude Tag on June 23, 2026, a research preview that embeds Claude directly into team chat workspaces instead of keeping it as a separate tool you have to open in another tab. The first platform is Slack, where Claude joins your workspace as a named team member you can mention just like a colleague.

The basic setup is simple: once added to a workspace, Claude appears as a participant. You can mention it directly in channels or threads. When you do, Claude can read and search through relevant Slack messages to ground its responses in what your team has actually been discussing — past decisions, inside terminology, project history — rather than relying only on what you type in a single message.

How this differs from what Slack already offers

Slack has had Claude integration before, allowing the model to pull in message context when you ask it something. Claude Tag is different because it treats Claude as a persistent member of your team rather than a tool you call in when needed. The architectural change is real: the model now has a standing identity within your collaboration space, not just a connection to it. Whether teams will actually use it more often as a result is something Anthropic will test during this research preview phase.

The practical questions ahead

Slack threads move fast, involve large teams, and often contain sensitive operational details. Before this becomes a standard feature, engineering and security teams will want clarity on how Claude Tag handles access — which messages it can read, whether that's limited to specific channels or roles, and how permissions work. That's not a minor detail when an AI model is sitting inside your internal communication layer.

Anthropic has not yet announced which other platforms will follow Slack, though the company frames Slack as the starting point rather than the final destination.

Broader context

AI assistants embedded in the collaboration tools where people actually work — rather than kept in their own separate chat boxes — has become the industry pattern. Microsoft's Copilot sits inside Teams and Office. Google's Gemini is integrated into Workspace. Anthropic's Claude Tag follows that same direction. The company has historically built Claude around nuanced reasoning and the ability to handle long context, so it will be worth watching whether those strengths show up differently inside a thread-based, back-and-forth environment compared to a direct conversation with the model.

In this author's view, the practical question for teams already using Claude is adoption friction. A Slack-native presence removes the friction of switching to another tab or copying and pasting thread content into a chat window. Historically, that kind of convenience has mattered more than raw capability improvements in determining whether a tool becomes a daily habit or an occasional lookup.

Claude Tag is available in research preview as of June 23, 2026, via Anthropic's announcement.